See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/379879007

Artificial Intelligence in Healthcare Delivery: Prospects and Pitfalls

Article in Journal of Medicine Surgery and Public Health · April 2024

DOI: 10.1016/j.glmedi.2024.100108

CITATIONS READS

9 223

6 authors, including:

David Olawade Aanuoluwapo Clement David-Olawade

University of East London University Hospitals Of Leicester NHS Trust
68 PUBLICATIONS 476 CITATIONS 17 PUBLICATIONS 81 CITATIONS

SEE PROFILE SEE PROFILE

Ojima Zechariah Wada Akinsola Asaolu

Hamad bin Khalifa University Derbyshire Healthcare NHS Foundation Trust
37 PUBLICATIONS 249 CITATIONS 3 PUBLICATIONS 12 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by David Olawade on 19 August 2024.

The user has requested enhancement of the downloaded file.

 Journal of Medicine, Surgery, and Public Health 3 (2024) 100108

Contents lists available at ScienceDirect

Journal of Medicine, Surgery, and Public Health

journal homepage: www.sciencedirect.com/journal/journal-of-medicine-surgery-and-public-health

Artificial intelligence in healthcare delivery: Prospects and pitfalls

David B. Olawade a, b, *, Aanuoluwapo C. David-Olawade c, Ojima Z. Wada d, Akinsola J. Asaolu e,
Temitope Adereni f, Jonathan Ling g
a
Department of Allied and Public Health, School of Health, Sport and Bioscience, University of East London, London, United Kingdom
b
Department of Public Health, York St John University, London, United Kingdom
c
Department of Nursing, University of Derby, Derby, United Kingdom
d
Division of Sustainable Development, College of Science and Engineering, Hamad Bin Khalifa University, Qatar Foundation, Doha, Qatar
e
Center for Research and Development, Kingsway Hospital, Derbyshire Healthcare NHS Foundation Trust, Derby, United Kingdom
f
Department of Public Health, University of Dundee, Dundee, United Kingdom
g
Faculty of Health Sciences and Wellbeing, University of Sunderland, United Kingdom

A R T I C L E I N F O A B S T R A C T

Keywords: This review provides a comprehensive examination of the integration of Artificial Intelligence (AI) into
Artificial intelligence healthcare, focusing on its transformative implications and challenges. Utilising a systematic search strategy
Healthcare across electronic databases such as PubMed, Scopus, Embase, and ScienceDirect, relevant peer-reviewed articles
Health equality
published in English between January 2010 till date were identified. Findings reveal AI’s significant impact on
Healthcare delivery
AI tools
healthcare delivery, including its role in enhancing diagnostic precision, enabling treatment personalisation,
facilitating predictive analytics, automating tasks, and driving robotics. AI algorithms demonstrate high accuracy
in analysing medical images for disease diagnosis and enable the creation of tailored treatment plans based on
patient data analysis. Predictive analytics identify high-risk patients for proactive interventions, while AI-
powered tools streamline workflows, improving efficiency and patient experience. Additionally, AI-driven ro­
botics automate tasks and enhance care delivery, particularly in rehabilitation and surgery. However, challenges
such as data quality, interpretability, bias, and regulatory frameworks must be addressed for responsible AI
implementation. Recommendations emphasise the need for robust ethical and legal frameworks, human-AI
collaboration, safety validation, education, and comprehensive regulation to ensure the ethical and effective
integration of AI in healthcare. This review provides valuable insights into AI’s transformative potential in
healthcare while advocating for responsible implementation to ensure patient safety and efficacy.

Introduction environment, and lifestyle affect their health by analysing vast volumes
of medical data [5,6]. Using this information, medical professionals can
Artificial intelligence (AI) is not a novel concept in the healthcare create personalised treatment programmes that are more effective.
sector and has already found widespread utilisation. Its integration is AI can also be applied to boost the effectiveness of healthcare sys­
increasingly prevalent in healthcare delivery, with the potential to tems. AI algorithms, for instance, can be used to extract patient medical
significantly enhance diagnostic precision, treatment efficacy, and records and pinpoint those most likely to develop specific ailments [7,
overall patient care [1]. The creation of diagnostic tools is one area in 8]. This can make it easier for healthcare practitioners to focus their
which AI is already utilised in healthcare. AI algorithms can be trained, resources to ensure patients receive the care they require. However,
for instance, to analyse medical images like X-rays and CT scans for some challenges must be addressed before AI can fully realise its po­
indications of disease or injury [2,3]. By doing so, medical professionals tential in the healthcare sector. The quality and quantity of training
may be able to make more accurate diagnoses, identify problems earlier data, for instance, may impact the precision of AI-based diagnostic tools
and therefore use less aggressive or more focused treatments [3,4]. The [9,10]. Applying AI in healthcare also raises ethical questions due to
creation of personalised medicine is another healthcare area where AI is potential biases in algorithms and the possibility of job loss for health­
employed. AI can assist doctors in understanding how patient genetics, care personnel [11–13]. Concerns also exist related to the acceptability

* Corresponding author at: Department of Allied and Public Health, School of Health, Sport and Bioscience, University of East London, London, United Kingdom.
E-mail address: olawadedavid@gmail.com (D.B. Olawade).

https://doi.org/10.1016/j.glmedi.2024.100108
Received 5 January 2024; Received in revised form 15 April 2024; Accepted 15 April 2024
Available online 16 April 2024
2949-916X/© 2024 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
D.B. Olawade et al. Journal of Medicine, Surgery, and Public Health 3 (2024) 100108

of AI-driven decision-making by healthcare practitioners [14]. to provide an up-to-date overview of the field. Articles were screened
Thus, this review embarks on an exploration of the integration of AI based on their relevance to the review objectives, with a focus on studies
within the healthcare domain, driven by the emergence of novel appli­ examining AI’s impact on diagnostic precision, treatment person­
cations and transformative implications. The rationale for delving into alisation, predictive analytics, automated tasks, and AI-driven robotics
this study stems from AI’s progressive integration into healthcare, in healthcare settings. Non-peer-reviewed articles and those unrelated to
signifying a paradigm shift with profound implications. The novelty of AI in healthcare were excluded from the review.
this review lies in its comprehensive analysis of AI’s multifaceted impact The systematic search strategy yielded a total of 500 articles across
on healthcare delivery, encompassing diagnostic precision, treatment the selected electronic databases, as follows:
personalisation, and healthcare system optimization. Furthermore, the PubMed: 120 articles
study aims to spotlight AI’s role in the creation of personalised medicine. Scopus: 100 articles
In addition, this review comprehensively examines recent advance­ Embase: 80 articles
ments, challenges, and ethical implications surrounding AI’s integration IEEE Xplore: 120 articles
into healthcare. By elucidating the multifaceted impact of AI, this ACM Digital Library: 80 articles
research aims to navigate the nuances of its adoption in healthcare, After initial screening based on title and abstract relevance, a subset
providing insights into its implications and potential future trajectories. of articles meeting the inclusion criteria underwent full-text review.
Ultimately, the study endeavours to shed light on the transformative Across all databases, a total of 300 articles were screened, resulting in
potential of AI in revolutionising healthcare delivery. the exclusion of 200 articles based on predefined criteria. Subsequently,
100 articles were selected for full-text review. Through rigorous
Methods screening, 80 high-quality, relevant articles were included in the review,
ensuring the incorporation of studies pertinent to the review objectives
This review employed a systematic approach to examine the current and enhancing the transparency and reproducibility of the review
state of AI applications in healthcare. The search strategy encompassed a process.
thorough exploration of electronic databases, including PubMed, Sco­
pus, Embase, IEEE Xplore, and ACM Digital Library, using a combination Results
of relevant search terms. The search terms utilised combinations of
keywords such as "artificial intelligence," "AI," "machine learning," "deep History of AI in healthcare
learning," "healthcare," "medical imaging," "predictive analytics," "ro­
botics," "virtual assistants," and "healthcare systems," among others. The use of AI in healthcare has a long history, going back to the early
Boolean operators (AND, OR) were used to refine the search and ensure years of AI research in the 1950s and 1960s [15] Early AI efforts
relevance to the review objectives. concentrated on creating expert systems, which were intended to imitate
Inclusion criteria comprised peer-reviewed articles published in the capability of human experts to make decisions [16]. These systems
English between January 2010 and the present, focusing on AI appli­ could analyse data and offer recommendations, but their capabilities
cations in healthcare. The rationale for selecting this timeframe lies in were constrained by the amount and quality of knowledge that could be
capturing the recent advancements and developments in AI technology stored in them [17].
and its applications in healthcare. The last decade has witnessed sig­ The availability of large volumes of medical data and developments
nificant progress in AI research and its integration into healthcare sys­ in computer technology fuelled a fresh interest in AI in healthcare in the
tems, making it crucial to focus on articles published during this period 1980s and 1990s [15,18] which led to the creation of novel AI

Fig. 1. PRISMA flowchart for selection of included studies.

2
D.B. Olawade et al. Journal of Medicine, Surgery, and Public Health 3 (2024) 100108

techniques more adaptable and potent than expert systems, like neural Table 1
networks and decision trees. Numerous applications, including decision Comprehensive Overview of AI Technologies and Tools in Diverse Healthcare
support systems and diagnostic tools [8,13], were created using these Applications.
principles, with the creation of an expert system for diagnosing breast AI Specific AI Use Case Example of tools
cancer being one of the first significant achievements of AI in healthcare. Application Techniques
The IDC-P (Intelligent Decision Support System for Breast Cancer Medical DL: • Segmenting and − Google’s DeepMind
Diagnosis) system, which included information from mammograms and Imaging Convolutional classifying Health: Enhances
other medical testing, successfully diagnosed breast cancer [19]. Analysis Neural tumors in image
[70,71] Networks radiology scans: segmentation.
Due to the development of improved AI techniques like deep learning
(CNNs), Canny Edge − Zebra Medical
and the accessibility of vast amounts of data, the use of AI in healthcare Generative Detection, K- Vision: Improves
has significantly increased in recent years. Numerous applications, Adversarial Means Clustering, image quality in
including diagnostic tools, predictive analytics, and personalised medi­ Networks Convolutional radiology.
cine [5,6], have been created using these principles. The creation of (GANs), &. Neural Networks
Holistically (CNNs) Aidoc: AI assistance
deep-learning algorithms to analyse medical images represents one of
−
Nested • Radiomics: Uses in detecting
the most remarkable achievements of AI in healthcare [1,20]. Networks (HNN) image features to abnormalities in X-
These algorithms have been utilised to create high-accuracy diag­ ML: Support assess disease rays.
nostic systems for various diseases, including cancer, heart disease and Vector Machines progression and
(SVM) treatment
eye disease. Future healthcare uses of AI are likely to be even more
response.
cutting-edge as the field of AI develops [13]. For example, Neuralink is Predictive Machine • Predicting the − SAS Healthcare
an emerging AI which functions by integrating a wireless brain chip into Analysis Learning risk of developing Analytics: Forecasts
the human brain, offering individuals with significant paralysis the [71,72] Algorithms: chronic diseases: disease risks.
capability to manipulate devices such as robotic limbs and smartphones SVM, XGBoost, Logistic
K-Nearest Regression, Optum360
through their thoughts [67]. However, there are also worries about bias
−
Neighbor, and Support Vector predictive
and errors in AI-based systems [21] and the need for thorough moni­ Naïve Bayes Machines analytics:
toring and evaluation. Nevertheless, these applications can improve (SVMs), Random Optimizes resource
patient outcomes and save costs [22]. As revealed by a recent review, Forests allocation.
• Anticipating
integrating AI in healthcare significantly reduces expenses when
patient
compared to conventional methods of determining health diagnosis, readmission
resulting in a time savings between 3.3 h per day to 15.2 h per day and rates: Time Series
cost savings ranging from 1,667 USD per day per hospital to 17,881 USD Analysis, ARIMA
per day per hospital [68]. Models, Neural
Networks
Overall, AI works by simulating human intelligence. It operates
Charting Natural • Summarizing − Nuance
through the collection, analysis, and interpretation of vast amounts of [73,74] Language patient records: Communications:
data, enabling it to learn from patterns and make decisions with minimal Processing Topic Modeling, Transcribes spoken
human intervention [69]. By harnessing algorithms and computational (NLP) Sentence words into EHR
Segmentation, text.
processes, AI systems can adapt and improve their performance over
Entity Extraction.
time. This self-improvement mechanism allows AI to achieve a level of • Generating − Cerner PowerChart:
precision and efficiency in tasks ranging from pattern recognition to clinical notes: Embeds AI support
predictive analysis. AI technologies such as Convolutional Neural Net­ Rule-based sys­ for clinical
works (CNNs) are utilised for medical imaging analysis, while Machine tems, Statistical documentation.
language models,
Learning Algorithms power predictive analysis in healthcare. Natural
Recurrent Neural
Language Processing (NLP) techniques are employed for charting pur­ Networks (RNNs)
poses, and AI-powered chatbots and virtual assistants facilitate Chatbots and AI-powered • Resolving patient − Ada Health: AI
communication and support within healthcare settings. Additionally, Virtual Chatbots and inquiries: chatbot for patient
Assistants Virtual Question symptom
AI-driven robots, integrating robotics and machine learning, are utilised
[75,76] Assistants answering assessment.
for various healthcare tasks. systems, Chatbots − HealthTap: Offers
Table 1 below outlines specific AI technologies and tools employed with rule-based AI-driven virtual
within distinct applications in the healthcare system. interactions assistants for medi­
• Scheduling cal information.
appointments:
Medical Imaging analysis Calendar − Woebot: AI chatbot
management designed for mental
For the diagnosis and treatment of a variety of illnesses, including systems, health support.
cancer, heart disease, and eye disease, medical imaging is a crucial tool. Appointment
booking chatbots
By automatically detecting and interpreting images, AI has the potential
AI-driven Robotics and • Surgical robots − Intuitive Surgical
to increase the precision and effectiveness of medical imaging analysis Robots Machine for minimally (da Vinci Surgical
[4,23]. Using deep learning algorithms - machine learning that employs [77,78] Learning invasive System): AI-
neural networks to evaluate massive volumes of data - for image analysis Integration procedures: Da powered surgical
is one of the most promising applications of AI in medical imaging [24]. Vinci Surgical assistance.
System, ZEPHYR Xenex: Uses AI for
These algorithms can be trained to spot patterns in images taken for
−
Surgical System disinfection robots
medical purposes and to detect or predict the presence of disease or • Drug delivery in healthcare
damage. robots for precise settings.
The creation of deep learning algorithms for the identification of skin dosing: Robotic
systems for Aethon:
cancer represents one of the most significant achievements of AI in −
intravenous drug Autonomous
medical imaging. Deep learning algorithms have been used to detect administration, robotic solutions
skin cancer with an accuracy level equivalent to that of human derma­ Insulin delivery for healthcare
tologists [25]. Algorithms have also been utilised to create diagnostic robots logistics.

3
D.B. Olawade et al. Journal of Medicine, Surgery, and Public Health 3 (2024) 100108

tools for breast and lung cancer [26]. Such algorithms can accurately possibility of biases in the algorithms as well as the requirement for
evaluate CT and MRI images to find tumours and other indications of continual monitoring and review [13]. In predictive analytics for med­
malignancy. ical care, especially concerning acute diseases, several general biases
AI-based algorithms may also increase the effectiveness of analysis of and limitations emerge with potential implications for patient care. The
medical imaging. For instance, regions of interest in images can be effectiveness of these analytics tools is often constrained by the diversity
automatically identified and marked using AI, which can reduce the and size of the datasets used to train them, which can limit their
time and labour needed for manual analysis [27]. AI-based algorithms applicability across different patient demographics and disease types
have also been used to evaluate retinal images to detect eye conditions [79,80]. A reliance on domain experts for selecting which clinical fea­
including glaucoma and diabetic retinopathy with great precision [28, tures to include may result in missing critical predictors due to over­
29]. looked complex factors, such as comorbidities and environmental
However, like all technology, there are several difficulties with AI- influences [79]. Furthermore, the complexity of deep learning models,
based medical imaging analysis. One of the greatest constraints is the while powerful, poses challenges in interpreting how these models make
necessity for a lot of high-quality data to train and validate the algo­ their predictions. This opacity can hinder clinicians’ trust in the tools, as
rithms. Concerns exist around the possibility of biases in the algorithms understanding the rationale behind recommendations is crucial for
as well as the requirement for continual monitoring and review [12–14]. clinical decisions [79,80]. Addressing these issues by enhancing data
The reliance on data-driven techniques necessitates high-quality, varied representation, expanding dataset diversity, and improving model
datasets to mitigate biases and enhance model performance [70]. interpretability is essential to developing more accurate, trustworthy,
However, challenges arise from inconsistent imaging protocols across and equitable predictive analytics in healthcare.
clinical settings, the vast array of clinical scenarios, and the unique
characteristics of patient cohorts in different clinics [71]. The organi­ Charting, chatbots and virtual assistants
sation and standardisation of data from diverse practices present sig­
nificant hurdles. Moreover, the critical task of data curation, including The process of charting involves entering patient data into electronic
accurate image labelling, is increasingly difficult due to the rapid growth health records (EHRs). AI-enabled charting solutions can evaluate pa­
in the volume of medical images and the expertise required for precise tient data and generate chart notes automatically, which can increase
annotation [71]. These issues are compounded by privacy concerns and the accuracy of patient records and completeness while also saving time.
regulations that restrict the sharing of patient images across institutions, Charting tools can be used to extract information from both organised
further complicating the development and application of AI in medical and unstructured data, such as lab results and free-text notes [35].
imaging [70,71]. The potential consequences of biases and limitations in Chatbots, often known as virtual assistants, are computer pro­
AI applications for medical imaging include compromised diagnostic grammes that can have voice or text conversations with users. They can
accuracy, the perpetuation of existing healthcare disparities, and a lack be used to inform patients, respond to their inquiries, and make ap­
of generalisability across diverse patient populations. Inaccurate or pointments. Virtual assistants can help with patient triage, medication
biased AI models can lead to misdiagnoses or overlooked conditions, management, education, and assistance in the field of healthcare [36].
particularly in underrepresented groups, thereby affecting patient out­ The creation of natural language processing (NLP) algorithms that can
comes. Additionally, the difficulties in data sharing due to privacy automatically extract information from unstructured data, like free-text
concerns and the variability in imaging protocols can hinder the notes, and structured data, like lab results, is one example of AI-powered
development of robust, universally applicable AI tools, potentially charting tools [34–36]. These algorithms can be used to find pertinent
slowing the progress towards more effective and personalised healthcare data, like diagnoses and treatment plans, and automatically add these to
solutions. a patient’s electronic health record [34].
Virtual assistants in the healthcare sector, such as Babylon Health,
Predictive analytics Ada, and Infermedica[37] can help patients by triaging their symptoms,
offering health information, and setting up appointments. Consequently,
By identifying those at risk of disease and delivering tailored in­ the adoption of AI-powered charting tools and virtual assistants has the
terventions, the use of AI for predictive analytics in healthcare offers the potential to improve patient experience while also increasing the effi­
potential to improve patient outcomes and save costs [30]. In order to ciency and accuracy of healthcare delivery. These technologies are still
forecast future health outcomes, predictive analytics analyses data from relatively new, so it remains crucial that constant monitoring and
electronic health records (EHRs), claims data, and other sources using evaluation are necessary to guarantee their efficacy. While these AI
statistical models and machine learning algorithms [31,32]. Risk strat­ applications are promising, they still have significant limitations to
ification is an important use of AI for predictive analytics in the overcome. For example, chatbots and virtual assistants in healthcare
healthcare industry. Risk stratification entails identifying people who face several challenges, including patient difficulties in using technology
are highly susceptible to developing a specific ailment and offering and a general lack of enthusiasm for following advice from digital
targeted therapies [33]. AI can, for instance, be used to pinpoint those platforms [75]. The impersonal nature of telemedicine may deter pa­
who are most at risk of acquiring diabetes, heart disease, or cancer and tients and doctors alike due to the reduced capacity for human
offer them individualised interventions like dietary adjustments, medi­ connection. Moreover, integrating voice technology into healthcare
cation management, or screenings [33]. necessitates further regulatory oversight, with a need for comprehensive
By identifying patients who have a high chance of being readmitted guidelines that address security, privacy, and the risk of cyberattacks, as
to the hospital, AI-based predictive analytics can also be utilised to these technologies can be vulnerable to hacking, especially in the
improve patient outcomes. AI can identify patients at high risk of context of smart home devices interacting with personal assistants [76].
readmission and offer targeted interventions to lower that risk by
evaluating data from EHRs and claims databases [34]. Through classi­ AI-driven robots
fying patients who are most likely to respond to particular therapies, AI
can also be used to increase the effectiveness of healthcare. AI can By automating repetitive processes and delivering targeted in­
pinpoint patients who are likely to respond to a certain treatment by terventions, the use of AI robots in healthcare offers the potential to
examining data from EHRs and claims data, which can help optimise the enhance patient outcomes and save costs. AI robots come in a variety of
use of resources and lower the cost of care [30]. For these applications to shapes and sizes, including exoskeletons, mobile robots, and humanoid
be effective, huge quantities of high-quality data must be available for robots. One of the primary areas where AI robots are being used in
the training and validation of algorithms. Concerns exist around the healthcare is rehabilitation [38,39]. Robots with AI can help patients

4
D.B. Olawade et al. Journal of Medicine, Surgery, and Public Health 3 (2024) 100108

with physical therapy exercises by offering individualised interventions inform drug discovery.
to enhance their functional and motor capabilities [39]. Specific examples of companies that provide healthcare-related AI
AI robots can also be employed to help with physically strenuous or technologies across North America, Africa, Europe, and Asia are pro­
precise tasks, such as tele-surgery, in which a surgeon performs surgery vided in Table 2.
using a robotic arm under remote control [40]. These robots can execute
surgical treatments with higher precision and flexibility than traditional Discussion
approaches, and they can be especially helpful for those that are chal­
lenging or impossible to accomplish with conventional surgical While AI holds great promise for improving healthcare delivery,
techniques. many challenges must be overcome before it can be broadly imple­
AI robots can also be utilised in the geriatric care sector to assist the mented. The absence of standardised and accessible data is a significant
elderly with activities of daily living, including dressing, grooming, and restriction. It can be challenging to transfer and analyse data across
bathing, as well as to provide companionship [41]. The use of AI robots multiple institutions because many healthcare organisations have their
for geriatric care is not widespread due to significant cost implications. own data storage systems and formats [60], and formal data sharing
“Eldercare robots” are more common in countries like Japan, with a agreements may not be in place. The challenge is more significant in
significant ageing population [78]. AI robots can also be employed to sub-Saharan Africa (SSA) and other under-developing/developing re­
help with jobs that are hazardous or challenging for humans to com­ gions, where the electronic collection of healthcare data is not routine
plete, including in contaminated environments or emergency response. [61]. Thus, creating AI technologies specific to SSA’s demographics
They may move through dangerous situations, evaluating the situation could pose a challenge. In addition to data availability and quality is­
and giving the emergency response team information in real-time [42]. sues, other hindrances in Africa are inadequate/unclear legal and policy
While robotics shows great potential to improve efficiency and frameworks and insufficient funding and infrastructure [62]. Health
reduce medical errors associated with routine procedures, there are bodies in SSA can remove these barriers by reviewing current digital
concerns about its cost implications. This is one of the major limitations health policies that can serve restriction to AI adoption. Since AI relies
of implementing (introducing and maintaining) such technologies in on accessibility to high-quality clinical data, these health organisations
countries like India [77]. In a study that examined the cost implications must intentionally create a regulatory framework to manage health data
of employing robotics in urology in Canada, the financial burden of and its collection. The advancement of AI technologies in developing
incorporating robot-assisted prostatectomy was substantial, with initial countries could increase the already existing global health inequality
acquisition costs exceeding $2 million USD and additional expenses gap.
related to operation, maintenance, and training [81,82]. The net cost of The inability to interpret AI models is another concern due to the
running a robotic surgery program can reach up to $3.5 million CAD high technical knowledge requirement of understanding machine
over seven years for prostatectomy alone, with each robot-assisted learning interpretability methods and its application to workflows, for
procedure costing significantly more than its open or laparoscopic instance, image classification, saliency maps and sensitivity analysis.
counterparts [81,82]. However, in the long term there are potential The multi-perspective nature of AI models makes them challenging to
reductions in incremental costs through increased procedural volume describe the decisions and how they were reached by deep learning
[82]. In addition, cost savings have also been reported in the short run. models as they can be challenging to interpret [63]. This can be espe­
For example, robot-assisted prostatectomy demonstrated early advan­ cially difficult in the healthcare industry, where transparency and trust
tages over open and laparoscopic surgeries in terms of perioperative are essential [64]. AI models may also be biased, especially if they were
outcomes, including shorter hospital stays, less blood loss, and lower developed with data unrepresentative of the population they would be
transfusion rates, as well as improved margins in prostate-confined tu­ applied to [11,12]. This may result in predictions and decisions that are
mours [81,83]. These benefits extend to better postoperative continence
and sexual function, which significantly impact patient quality of life
post-surgery [84]. Table 2
By drastically reducing the time and expense needed to produce new Some healthcare AI-companies and their services.
pharmaceuticals, AI also has the potential to transform the drug devel­ Company AI Function Year and locations Reference
opment process. A variety of data sources, including scientific publica­ commissioned
tions, clinical trials, and chemical databases, can be analysed using AI Viz.ai Aneurysm detection 2018, [48]
algorithms to identify prospective drug targets and predict the effec­ software USA
tiveness and safety of novel compounds. Using AI-based algorithms to PathAI Precision pathology 2016 [49]
USA
conduct computational chemistry enables the process of finding new
Buoy Symptom assessment 2014 [50]
drugs to be greatly accelerated, through anticipating the characteristics chatbox USA
of micro molecules, such as their solubility, stability, and bioactivity Beth Israel AI-powered microscope 2017 [51]
[43]. Virtual screening is a different area where AI is being used in the Medical to detect blood infection USA
drug discovery process. Large databases of compounds can be examined Center
Iterative Health Precision 2017 [52]
using AI-based algorithms to identify potential drug candidates that are gastroenterology USA
highly likely to bind to a particular protein target [44,45]. VirtuSense Movement anomaly 2012 [53]
In order to find potential drug targets for a variety of diseases, AI may detection USA
also be used to analyse vast amounts of genetic data. For instance, cancer RxAll Drug authentication 2016 [54]
Africa
patient genomic data can be explored using AI-based algorithms to
Viebeg Healthcare facilities 2018 [55]
identify new therapeutic targets for customised therapy [46]. Addi­ Technologies procurement Rwanda
tionally, by examining data from clinical trials and electronic health Lunit Detection of 2013 [56]
records, AI can be used to detect probable negative effects of medica­ abnormalities in chest South Korea
tions (EHRs). AI can help increase the safety of new pharmaceuticals and radiographs
Promaton Dental diagnostics 2017 [57]
lower the likelihood of adverse events by identifying potential side ef­ Netherlands
fects early in the drug development process [47]. The use of AI to Aidence Lung cancer detection 2017 [58]
identify new compounds and rule out others is based on the quality of via precise imaging Netherlands
the data available. Improving data quality and sharing data between Chronolife Health monitoring and 2019 [59]
prediction France
databases is of critical importance in optimizing the algorithms used to

5
D.B. Olawade et al. Journal of Medicine, Surgery, and Public Health 3 (2024) 100108

unfair or inaccurate. frameworks governing AI in healthcare. This involves in-depth

Additionally, some AI models may lack robustness, making them exploration of region-specific policies, guidelines, and regulations
susceptible to manipulation by hostile instances and less likely to to ensure the responsible and ethical use of AI technologies. Studies
generalise effectively to new data [65,85]. The absence of rules and should focus on identifying potential biases within algorithms and
regulations for AI in healthcare is also a concern. As the use of AI within establishing protocols for transparency, fairness, and accountability
healthcare is still relatively new, there are no clear guidelines for its in AI-driven decision-making processes.
creation, distribution and application. This could result in a lack of 2. Human-AI Collaboration: Research efforts should delve into opti­
clarity and uncertainty over how to effectively deploy AI in this context mizing the collaboration between AI technologies and healthcare
[13]. professionals. This involves investigating strategies for seamless
While these are potential downsides, these challenges can be miti­ integration of AI tools into existing healthcare workflows, fostering
gated by proper regulation, design, and implementation of AI systems, trust, and ensuring effective communication between AI systems and
as well as by involving healthcare professionals, ethicists and other human caregivers. Studies can explore training programs that enable
stakeholders in the development process [66]. To overcome these re­ healthcare professionals to understand and effectively utilize AI-
strictions and guarantee that AI is created and applied in a responsible generated insights while retaining the pivotal role of human judg­
and ethical manner, more research is required. ment in patient care.
Addressing the limitations and challenges of AI in healthcare re­ 3. Safety and Effectiveness: Emphasis should be placed on continual
quires multifaceted solutions and ongoing research efforts. One critical assessment and validation of AI tools in healthcare settings to ensure
challenge is the lack of standardized and accessible data, especially in their safety, efficacy, and accuracy. Longitudinal studies and real-
regions like sub-Saharan Africa where electronic healthcare data world applications are pivotal to monitor the performance of AI al­
collection is not routine [61]. To overcome this, health bodies in these gorithms, identify potential shortcomings, and refine these tools over
regions can review and establish digital health policies that facilitate AI time. Research should focus on developing standardized protocols
adoption and ensure proper management of health data. Additionally, for evaluating and validating AI technologies in clinical settings.
efforts should focus on developing AI technologies tailored to the de­ 4. Education and Training: Future research should prioritize education
mographics and healthcare systems of these regions. Another concern is and training initiatives for healthcare professionals and stakeholders.
the interpretability of AI models, which requires high technical knowl­ This involves developing educational programs to enhance under­
edge and can be particularly challenging in healthcare where trans­ standing, acceptance, and proficiency in utilizing AI tools in
parency and trust are crucial. Addressing this challenge involves healthcare. Addressing the concerns and knowledge gaps surround­
developing interpretable AI methods and involving healthcare pro­ ing AI technologies among healthcare professionals is vital for their
fessionals in the interpretation process. Furthermore, mitigating bias in successful adoption and integration into clinical practice.
AI models and ensuring their robustness against manipulation are 5. Regulation and Governance: Research efforts should advocate for
essential steps in improving AI’s reliability and fairness in healthcare. comprehensive regulatory frameworks and governance mechanisms
Establishing clear rules and regulations for AI in healthcare is also specific to AI in healthcare. Collaboration between policymakers,
necessary to provide guidance on its creation, distribution, and appli­ regulatory bodies, technologists, and healthcare stakeholders is
cation. Ultimately, overcoming these challenges and ensuring respon­ crucial for formulating guidelines that balance innovation and
sible and ethical AI implementation requires collaboration among regulation, ensuring the ethical deployment and ongoing oversight
stakeholders and continued research efforts in the field. of AI technologies in healthcare.
Table 3 provides a comprehensive overview of the advantages and
disadvantages of AI in healthcare systems, explores potential users, and Future research should focus on interdisciplinary collaboration
highlights the implications associated with the integration of AI tech­ encompassing ethical, legal, technical and sociocultural dimensions. By
nologies in healthcare delivery. addressing these recommendations, researchers can pave the way for the
Recommendations for the future direction of research in this field responsible and effective integration of AI tools in healthcare, while
can be delineated into several key aspects, such as: ensuring ethical, safe and trustworthy implementation of these trans­
formative technologies. By embracing AI responsibly and proactively,
1. Ethical and Legal Frameworks: Future research should prioritize the healthcare systems can pave the way for a future where technology-
development and implementation of robust ethical and legal driven innovations drive improvements in patient outcomes and

Table 3
Advantages, Disadvantages, Users, and Implications of AI in Healthcare Systems.
Feature Advantages Disadvantages Potential Users Implications

Enhanced Diagnosis Accurately identify rare diseases and complex May require large amounts of training Physicians, specialists, Improve diagnostic accuracy and
and Treatment medical conditions data, which may not always be pathologists reduce misdiagnosis
[86] available
Personalised Tailor treatment plans to each patient’s unique Potential oversight of the significance Patients, caregivers Enhance patient engagement and
Healthcare [87] needs and preferences of patient-provider interaction and improve adherence to treatment
shared decision-making plans
Predictive Analytics Identify patients at high risk of developing Accuracy of predictive models may vary Physicians, hospitals, Proactively intervene to prevent
[88] chronic conditions or experiencing adverse depending on the data used and the health systems adverse events and improve
events complexity of the prediction population health
Automated Tasks Free up healthcare workers from routine tasks, May lead to job displacement and skill Hospital staff, Streamline workflows, improve
[89] allowing them to focus on more complex obsolescence among certain healthcare administrative efficiency, and reduce the burden
clinical duties professionals personnel on healthcare providers
Chatbots and Provide 24/7 access to healthcare information May not always be able to understand Patients, caregivers Improve patient satisfaction, reduce
Virtual Assistants and support, including appointment complex medical queries or provide wait times, and enhance patient
[85] scheduling, medication reminders, and adequate support for emergencies engagement
symptom tracking
AI-Driven Robots Perform minimally invasive procedures with May require significant upfront Surgeons, pharmacists, Revolutionize healthcare delivery,
[90] greater precision and dexterity investment and ongoing maintenance telemedicine providers reduce surgical complications, and
improve patient recovery

6
D.B. Olawade et al. Journal of Medicine, Surgery, and Public Health 3 (2024) 100108

population health. [4] O. Oren, B.J. Gersh, D.L. Bhatt, Artificial intelligence in medical imaging: switching
from radiographic pathological data to clinically meaningful endpoints, Lancet
Digit. Health 2 (9) (2020) e486–e488, https://doi.org/10.1016/S2589-7500(20)
Implications of the review 30160-6.
[5] J. Bajwa, U. Munir, A. Nori, B. Williams, Artificial intelligence in healthcare:
transforming the practice of medicine, Future Healthc. J. 8 (2) (2021) e188–e194,
The findings of this review underscore the growing importance of https://doi.org/10.7861/fhj.2021-0095.
integrating AI technologies into clinical practice. Healthcare pro­ [6] K.B. Johnson, W. Wei, D. Weeraratne, M.E. Frisse, K. Misulis, K. Rhee K, et al.,
fessionals need to familiarize themselves with these tools to harness Precision medicine, AI, and the future of personalized health care, Clin. Transl. Sci.
14 (1) (2020) 86–93, https://doi.org/10.1111/cts.12884.
their potential effectively. Training programs and continuing education [7] E.J. Topol, High-performance medicine: the convergence of human and artificial
initiatives should be developed to ensure healthcare providers are intelligence, Nat. Med. 25 (1) (2019) 44–56, https://doi.org/10.1038/s41591-018-
equipped with the necessary skills to leverage AI for improved patient 0300-7.
[8] V. Baxi, R. Edwards, M. Montalto, S. Saha, Digital pathology and artificial
care. This review highlights the need for further research into the intelligence in translational medicine and clinical practice, Mod. Pathol. 35 (2021)
development and validation of AI algorithms for various healthcare 22–32, https://doi.org/10.1038/s41379-021-00919-2.
applications. Future studies should focus on refining AI models, [9] S. Jayakumar, V. Sounderajah, P. Normahani, L. Harling, S.R. Markar,
H. Ashrafian, et al., Quality assessment standards in artificial intelligence
addressing issues related to bias and interpretability, and evaluating the
diagnostic accuracy systematic reviews: a meta-research study, NPJ Digit. Med. 5
long-term impact of AI interventions on patient outcomes. Collaborative (2022) 11, https://doi.org/10.1038/s41746-021-00544-y.
research efforts between academia, industry, and healthcare institutions [10] C.J. Kelly, A. Karthikesalingam, M. Suleyman, G. Corrado, D. King, Key challenges
are essential to advance the field of AI in healthcare. for delivering clinical impact with artificial intelligence, BMC Med. 17 (2019) 195,
https://doi.org/10.1186/s12916-019-1426-2.
AI has the potential to revolutionize preventive healthcare by [11] N. Naik, B.M.Z. Hameed, D.K. Shetty, D. Swain, M. Shah, R. Paul, et al., Legal and
enabling more accurate risk stratification, early detection of diseases, ethical consideration in artificial intelligence in healthcare: who takes
and personalized interventions. Policymakers and healthcare leaders responsibility? Front. Surg. 9 (2022) 862322 https://doi.org/10.3389/
fsurg.2022.862322.
should prioritize investments in AI-driven preventive strategies, such as [12] B. Khan, H. Fatima, A. Qureshi, S. Kumar, A. Hanan, J. Hussain, et al., Drawbacks
predictive analytics and population health management tools, to miti­ of artificial intelligence and their potential solutions in the healthcare sector,
gate the burden of chronic diseases and improve population health Biomed. Mater. Devices (2023) 1–8, https://doi.org/10.1007/s44174-023-00063-
2.
outcomes. The adoption of AI in healthcare necessitates the develop­ [13] F. Jiang, Y. Jiang, H. Zhi, Y. Dong, H. Li, S. Ma, et al., Artificial intelligence in
ment of robust regulatory frameworks to ensure patient safety, privacy healthcare: past, present and future, Stroke Vasc. Neurol. 2 (4) (2017) 230–243,
protection, and ethical use of data. Policymakers must work collabora­ https://doi.org/10.1136/svn-2017-000101.
[14] S. Prakash, J.N. Balaji, A. Joshi, K.M. Surapaneni, Ethical conundrums in the
tively with stakeholders to establish guidelines for the design, imple­ application of artificial intelligence (AI) in healthcare—a scoping review of
mentation, and evaluation of AI technologies in healthcare. reviews, J. Pers. Med. 12 (11) (2022) 1914, https://doi.org/10.3390/
Additionally, policies should address issues related to data sharing, jpm12111914.
[15] C.A. Kulikowski, Beginnings of artificial intelligence in medicine (AIM):
interoperability, and reimbursement to facilitate the widespread adop­
computational artifice assisting scientific inquiry and clinical art – with reflections
tion of AI solutions.Top of Form on present AIM Challenges, Yearb. Med. Inform. 28 (1) (2019) 249–256, https://
doi.org/10.1055/s-0039-1677895.
[16] J. Fox, Expert Systems and Theories of Knowledge, in: M.A. Boden (Ed.), In
Conclusion Handbook of Perception and Cognition, Academic Press, 1996, pp. 157–181,
https://doi.org/10.1016/B978-012161964-0/50008-X.
In conclusion, the integration of AI tools into healthcare holds [17] E.H. Shortliffe, Computer-Based Medical Consultations: MYCIN, Annals of Internal
Medicine [Internet], (1976) [cited 2023 Feb 17];85(6):831. Available from: 〈https
immense potential for enhancing patient outcomes, streamlining pro­ ://www.acpjournals.org/doi/10.7326/0003-4819-85-6-831_1〉.
cesses, and driving cost efficiencies. However, it is imperative to [18] I.H. Sarker, Deep learning: a comprehensive overview on techniques, taxonomy,
recognize that AI should complement rather than replace human applications and research directions, SN Comput. Sci. 2 (420) (2021), https://doi.
org/10.1007/s42979-021-00815-1.
expertise and decision-making in healthcare settings. Moreover, the [19] S.D. Roy, S. Das, D. Kar, F. Schwenker, R. Sarkar, Computer aided breast cancer
ethical and legal implications surrounding the deployment of AI tech­ detection using ensembling of texture and statistical image features, Sensors 21
nologies demand careful consideration, necessitating the development (11) (2021) 3628, https://doi.org/10.3390/s21113628.
[20] A.S. Ahuja, The impact of artificial intelligence in medicine on the future role of the
of region-specific policy frameworks. While AI offers unprecedented
physician, PeerJ 7 (2019) e7702, https://doi.org/10.7717/peerj.7702.
opportunities, its implementation should proceed cautiously, guided by [21] R. Challen, J. Denny, M. Pitt, L. Gompels, T. Edwards, K. Tsaneva-Atanasova,
robust regulation to safeguard against potential risks and ensure the Artificial intelligence, bias and clinical safety, BMJ Qual. Saf. 28 (3) (2019)
231–237, https://doi.org/10.1136/bmjqs-2018-008370.
ethical use of these transformative technologies. By striking a balance
[22] J. Wolff, J. Pauling, A. Keck, J. Baumbach, The economic impact studies of
between innovation and regulation, healthcare stakeholders can harness artificial intelligence in health care: systemic review, J. Med. Internet Res. 22 (2)
the full benefits of AI while upholding patient safety, efficacy, and (2020) e16866, https://doi.org/10.2196/16866.
ethical standards. [23] X. Tang, The role of artificial intelligence in medical imaging research, BJR|Open 2
(1) (2020) 20190031, https://doi.org/10.1259/bjro.20190031.
[24] L. Manco, N. Maffei, S. Strolin, S. Vichi, L. Bottazzi, L. Strigari, Basic of machine
learning and deep learning in imaging for medical physicists, Phys. Med. 83 (2021)
Declaration of competing interest 194–205, https://doi.org/10.1016/j.ejmp.2021.03.026.
[25] A. Esteva, B. Kuprel, R.A. Novoa, J. Ko, S.M. Swetter, H.M. Blau, et al.,
Dermatologist-level classification of skin cancer with deep neural networks, Nature
The authors declare that they have no known competing financial 542 (2017) 115–118, https://doi.org/10.1038/nature21056.
interests or personal relationships that could have appeared to influence [26] J.L. Espinoza, L.T. Dong, Artificial intelligence tools for refining lung cancer
the work reported in this paper. screening, J. Clin. Med. 9 (12) (2020) 3860, https://doi.org/10.3390/jcm9123860.
[27] Y. Gang, X. Chen, H. Li, H. Wang, J. Li, Y. Guo, et al., A comparison between
manual and artificial intelligence–based automatic positioning in CT imaging for
References COVID-19 patients, Eur. Radiol. 31 (2021) 6049–6058, https://doi.org/10.1007/
s00330-020-07629-4.
[28] S.K. Padhy, B. Takkar, R. Chawla, A. Kumar, Artificial intelligence in diabetic
[1] T. Davenport, T., R. Kalakota, The potential for artificial intelligence in healthcare,
retinopathy: a natural step to the future, Indian J. Ophthalmol. 67 (7) (2019)
Future Healthc. J. 6 (2) (2019) 94–98, https://doi.org/10.7861/futurehosp.6-2-94.
1004–1009, https://doi.org/10.4103/ijo.IJO_1989_18.
[2] T.A. Soomro, L. Zheng, A.J. Afifi, A. Ali, M. Yin, J. Gao, Artificial intelligence (AI)
[29] X. Huang, H. Wang, C. She, J. Feng, X. Liu, X. Hu, et al., Artificial intelligence
for medical imaging to combat coronavirus disease (COVID-19): a detailed review
promotes the diagnosis and screening of diabetic retinopathy, Front. Endocrinol.
with direction for future research, Artif. Intell. Rev. 52 (2) (2021) 1409–1439,
13 (2022) 946915, https://doi.org/10.3389/fendo.2022.946915.
https://doi.org/10.1007/s10462-021-09985-z.
[30] Erin McNemar, Adopting AI to Improve Patient Outcomes, Cost Savings, Health
[3] Y. Kumar, A. Koul, R. Singla, M.F. Ijaz, Artificial intelligence in disease diagnosis: a
Equality [Internet], HealthITAnalytics (2021) [cited 2023 Feb 17]. Available from:
systematic literature review, synthesizing framework and future research agenda,
〈https://healthitanalytics.com/news/adopting-ai-to-improve-patient-outcomes-c
J. Ambient Intell. Humaniz. Comput. (2022) 1–28, https://doi.org/10.1007/
ost-savings-health-equality〉.
s12652-021-03612-z.

7
D.B. Olawade et al. Journal of Medicine, Surgery, and Public Health 3 (2024) 100108

[31] J. Wong, M. Murray Horwitz, L. Zhou, S. Toh, Using machine learning to identify [58] Abidance, AI-powered clinical applications for the lung cancer pathway [Internet]
health outcomes from electronic health record data, Curr. Epidemiol. Rep. 5 (4) Aidence (2023) [cited 2023 Feb 18]. Available from: 〈https://www.aidence.com/〉.
(2018) 331–342, https://doi.org/10.1007/s40471-018-0165-9. [59] Chronolife, (2019). Available at: 〈https://www.chronolife.net/〉 [Accessed 17 Feb.
[32] S. Dash, S.K. Shakyawar, M. Sharma, S. Kaushik, Big data in healthcare: 2023].
management, analysis and future prospects, J. Big Data 6 (2019) 54, https://doi. [60] A. Rajkomar, E. Oren, K. Chen, A.M. Dai, N. Hajaj, M. Hardt, et al., Scalable and
org/10.1186/s40537-019-0217-0. accurate deep learning with electronic health records, NPJ Digit. Med. 1 (18)
[33] S. Singh, R. Zeltser, Cardiac Risk Stratification [Internet]. PubMed. Treasure Island (2018), https://doi.org/10.1038/s41746-018-0029-1.
(FL), StatPearls Publishing, 2020. 〈https://www.ncbi.nlm.nih.gov/books/NBK [61] F.F. Odekunle, R.O. Odekunle, S. Shankar S, Why sub-Saharan Africa lags in
507785/〉 (Available from). electronic health record adoption and possible strategies to increase its adoption in
[34] A. Ozonoff, C.E. Milliren, K. Fournier, J. Welcher, A. Landschaft, M. Samnaliev, et this region, Int. J. Health Sci. [Internet] 11 (4) (2017) 59–64. 〈https://www.ncbi.
al., Electronic surveillance of patient safety events using natural language nlm.nih.gov/pmc/articles/PMC5654179/〉.
processing, Health Inform. J. 28 (4) (2022), https://doi.org/10.1177/ [62] A. Owoyemi, J. Owoyemi, A. Osiyemi, A., A. Boyd, Artificial intelligence for
14604582221132429. healthcare in Africa, Front. Digit. Health 2 (6) (2020), https://doi.org/10.3389/
[35] C.J. Harrison, C.J. Sidey-Gibbons, Machine learning in medicine: a practical fdgth.2020.00006.
introduction to natural language processing, BMC Med. Res. Methodol. 21 (2021) [63] P. Linardatos, V. Papastefanopoulos, S. Kotsiantis, Explainable AI: a review of
158, https://doi.org/10.1186/s12874-021-01347-1. machine learning interpretability methods, Entropy 23 (1) (2020) 18, https://doi.
[36] T. Jadczyk, W. Wojakowski, M. Tendera, T.D. Henry, G. Egnaczyk, S. Shreenivas, org/10.3390/e23010018.
Artificial intelligence can improve patient management at the time of a pandemic: [64] A. Kiseleva, D. Kotzinos, P. De Hert, Transparency of AI in healthcare as a
the role of voice technology, J. Med. Internet Res. 23 (5) (2021) e22959, https:// multilayered system of accountabilities: between legal requirements and technical
doi.org/10.2196/22959. limitations, Front. Artif. Intell. 5 (2022) 879603, https://doi.org/10.3389/
[37] BIS Research, Major Trends Influencing the Healthcare Chatbots Market [Internet], frai.2022.879603.
blog.marketresearch.com, (2019) [cited 2023 Feb 18]. Available from: 〈https://blo [65] T. Chen, C. Guestrin, XGBoost: A Scalable Tree Boosting System. In: Proceedings of
g.marketresearch.com/major-trends-influencing-the-healthcare-chatbots-market〉. the 22nd ACM SIGKDD International Conference on Knowledge Discovery and
[38] F. Yuan, E. Klavon, Z. Liu, R.P. Lopez, X. Zhao, A systematic review of robotic Data Mining - KDD ’16. Association for Computing Machinery, New York, NY, USA,
rehabilitation for cognitive training, Front. Robot. AI 8 (2021) 605715, https://doi. (2016) 785–94. DOI:〈10.1145/2939672.2939785〉.
org/10.3389/frobt.2021.605715. [66] L. Petersson, I. Larsson, J.M. Nygren, P. Nilsen, M. Neher, J.E. Reed, et al.,
[39] M. Asada, K. Hosoda, Y. Kuniyoshi, H. Ishiguro, T. Inui, Y. Yoshikawa, et al., Challenges to implementing artificial intelligence in healthcare: a qualitative
Cognitive developmental robotics, A Surv., IEEE Trans. Auton. Ment. Dev. 1 (1) interview study with healthcare leaders in Sweden, BMC Health Serv. Res. 22
(2009) 12–34, https://doi.org/10.1109/TAMD.2009.2021702. (2022) 850, https://doi.org/10.1186/s12913-022-08215-8.
[40] A. Hamed, S.C. Tang, H. Ren, A. Squires, C. Payne, K. Masamune, et al., Advances [67] N. Agrawal, Neuralink: linking AI with the human mind, Lambert Post (2024). 〈htt
in haptics, tactile sensing, and manipulation for robot-assisted minimally invasive ps://thelambertpost.com/news/neuralink-linking-ai-with-the-human-mind/〉
surgery, noninvasive surgery, and diagnosis, J. Robot. 2012 (2012) 1–14, https:// (Available at).
doi.org/10.1155/2012/412816. [68] N.N. Khanna, M.A. Maindarkar, V. Viswanathan, J.F.E. Fernandes, S. Paul, M.
[41] R. Kachouie, S. Sedighadeli, R. Khosla, M.-T. Chu, Socially assistive robots in M. Bhagawati, et al., Economics of artificial intelligence in healthcare: diagnosis vs.
elderly care: a mixed-method systematic literature review, Int. J. Hum. - Comput. treatment, Healthcare (Basel) 10 (12) (2022) 2493, https://doi.org/10.3390/
Interact. 30 (5) (2014) 369–393, https://doi.org/10.1080/ healthcare10122493. PMID: 36554017; PMCID: PMC9777836.
10447318.2013.873278. [69] Y. Xu, X. Liu, X. Cao, C. Huang, E. Liu, S. Qian, X. Liu, Y. Wu, et al., Artificial
[42] C. Kyrkou, P. Kolios, T. Theocharides, M. Polycarpou, Machine learning for intelligence: a powerful paradigm for scientific research, Innovation (Camb.) 2 (4)
emergency management: a survey and future outlook, Proc. IEEE 111 (1) (2023) (2021) 100179, https://doi.org/10.1016/j.xinn.2021.100179. PMID: 34877560;
19–41, https://doi.org/10.1109/JPROC.2022.3223186. PMCID: PMC8633405.
[43] Y. Zhang, M. Luo, P. Wu, S. Wu, T.-Y. Lee, C. Bai, Application of computational [70] L. Pinto-Coelho, How artificial intelligence is shaping medical imaging technology:
biology and artificial intelligence in drug design, Int. J. Mol. Sci. 23 (21) (2022) a survey of innovations and applications, Bioengineering (Basel) 10 (12) (2023)
13568, https://doi.org/10.3390/ijms232113568. 1435, https://doi.org/10.3390/bioengineering10121435.
[44] P. Carracedo-Reboredo, J. Liñares-Blanco, N. Rodríguez-Fernández, F. Cedrón, F. [71] X. Tang, The role of artificial intelligence in medical imaging research, BJR Open 2
J. Novoa, A. Carballal, et al., A review on machine learning approaches and trends (1) (2019) 20190031, https://doi.org/10.1259/bjro.20190031.
in drug discovery, Comput. Struct. Biotechnol. J. 19 (2021) 4538–4558, https:// [72] N. Ghaffar Nia, E. Kaplanoglu, A. Nasab, Evaluation of artificial intelligence
doi.org/10.1016/j.csbj.2021.08.011. techniques in disease diagnosis and prediction, Discov. Artif. Intell. 3 (1) (2023) 5,
[45] M. Batool, B. Ahmad, S. Choi, A structure-based drug discovery paradigm, Int. J. https://doi.org/10.1007/s44163-023-00049-5.
Mol. Sci. 20 (11) (2019) 2783, https://doi.org/10.3390/ijms20112783. [73] K. Agatstein, Chart review is dead; long live chart review: how artificial
[46] C.G. Fischer, A. Pallavajjala, L. Jiang, V. Anagnostou, J. Tao, E. Adams, et al., intelligence will make human review of medical records obsolete, one day, Popul
Artificial intelligence-assisted serial analysis of clinical cancer genomics data Health Manag 26 (6) (2023) 438–440, https://doi.org/10.1089/pop.2023.0227.
identifies changing treatment recommendations and therapeutic targets, Clin. Epub 2023 Oct 4. PMID: 37792424; PMCID: PMC10698762. 73.
Cancer Res. 28 (11) (2022) 2361–2372, https://doi.org/10.1158/1078-0432.CCR- [74] Y. Wen, X. Li, F. Zeng, J. Lei, S. Chen, 基于人工智能的病历质控系统的应用研究
21-4061. [Application of Medical Record Quality Control System Based on Artificial
[47] A.O. Basile, A. Yahi, N.P. Tatonetti, Artificial intelligence for drug toxicity and Intelligence], 54 (6) (2023) 1263-1268. Chinese. doi: 〈10.12182/20231160206〉.
safety, Trends Pharmacol. Sci. 40 (9) (2019) 624–635, https://doi.org/10.1016/j. PMID: 38162053; PMCID: PMC10752767. 74.
tips.2019.07.005. [75] T. Jadczyk, W. Wojakowski, M. Tendera, T.D. Henry, G. Egnaczyk, S. Shreenivas,
[48] Viz, AI-powered care coordination platform for disease detection and workflow Artificial intelligence can improve patient management at the time of a pandemic:
optimization [Internet]. 〈www.viz.ai〉. (2022) [cited 2023 Feb 18]. Available from: the role of voice technology, J. Med. Internet Res. 23 (5) (2021) e22959, https://
〈https://www.viz.ai/〉. doi.org/10.2196/22959. PMID: 33999834; PMCID: PMC8153030. 75.
[49] PathAi, Improving Patient Outcomes with AI-Powered Pathology [Internet]. [76] S.A. Alowais, S.S. Alghamdi, N. Alsuhebany, T. Alqahtani, A.I. Alshaya, et al.,
〈www.pathai.com〉. (2023) [cited 2023 Feb 18]. Available from: 〈https://www. Revolutionizing healthcare: the role of artificial intelligence in clinical practice,
pathai.com/〉. BMC Med. Educ. 23 (1) (2023) 689, https://doi.org/10.1186/s12909-023-04698-
[50] Buoy, Buoy Health: Check Symptoms & Find the Right Care [Internet]. 〈www. z. PMID: 37740191; PMCID: PMC10517477. 76.
buoyhealth.com〉. (2023). Available from: 〈https://www.buoyhealth.com/〉. [77] N. Deo, A. Anjankar, Artificial intelligence with robotics in healthcare: a narrative
[51] S. Daley, Updated by Rose Velazquez [2023 Jan 3], 46 AI in Healthcare Examples review of its viability in India, Cureus 15 (5) (2023) e39416, https://doi.org/
Improving the Future of Medicine [Internet]. Built In. (2018) [cited 2023 Feb 18]. 10.7759/cureus.39416. PMID: 37362504; PMCID: PMC10287569.
Available from: 〈https://builtin.com/artificial-intelligence/artificial-intelligence-h [78] K. Denecke, C.R. Baudoin, A review of artificial intelligence and robotics in
ealthcare〉. transformed health ecosystems, Front. Med. 9 (2022), https://doi.org/10.3389/
[52] Iterative Health, Iterative Health – Bringing world-class GI care and treatment to fmed.2022.795957.
patients around the world [Internet]. iterative.health. (2023) [cited 2023 Feb 18]. [79] J.Q. Sheng, P.J. Hu, X. Liu, T.S. Huang, Y.H. Chen, Predictive analytics for care and
Available from: 〈https://iterative.health/〉. management of patients with acute diseases: deep learning-based method to
[53] VirtuSense, VirtuSense Technologies | Fall Prevention and RPM with AI [Internet]. predict crucial complication phenotypes, J. Med. Internet Res. 23 (2) (2021)
〈www.virtusense.ai〉. (2023) [cited 2023 Feb 18]. Available from: 〈https://www. e18372, https://doi.org/10.2196/18372. PMID: 33576744; PMCID: PMC7910123.
virtusense.ai/〉. [80] B. Van Calster, L. Wynants, D. Timmerman, E.W. Steyerberg, G.S. Collins,
[54] RxAll, Digital infrastructure for healthcare [Internet]. RxAll - Digital infrastructure Predictive analytics in health care: how can we know it works? J. Am. Med. Inf.
for healthcare (2023) [cited 2023 Feb 18]. Available from: 〈https://rxall.net/〉. Assoc. 26 (12) (2019) 1651–1654, https://doi.org/10.1093/jamia/ocz130. PMID:
[55] African Development Bank, How Rwanda is using Artificial Intelligence to improve 31373357; PMCID: PMC6857503.
healthcare [Internet]. afdb.org. (2022) [cited 2023 Feb 18]. Available from: [81] C. Ho, E. Tsakonas, K. Tran, et al., Robot-Assisted Surgery Compared with Open
〈https://www.afdb.org/en/success-stories/how-rwanda-using-artificial-inte Surgery and Laparoscopic Surgery: Clinical Effectiveness and Economic Analyses
lligence-improve-healthcare-55309〉. [Internet]. Ottawa, ON: Canadian Agency forDrugs and Technologies in Health
[56] Lunit, Conquer Cancer through AI [Internet]. 〈www.lunit.io〉. (2021) [cited 2023 (2011) Sept (CADTH TechnologyReport, No. 137) [cited 2024 Mar 23] Available
Feb 18]. Available from: 〈https://www.lunit.io/en〉. from: 〈http://www.cadth.ca/en/products/health-technology-assessment/publicati
[57] Promaton, Artificial intelligence for the dental industry [Internet]. 〈www.proma on/2682〉.
ton.com〉. (2021) [cited 2023 Feb 18]. Available from: 〈https://www.promaton.
com/〉.

8
D.B. Olawade et al. Journal of Medicine, Surgery, and Public Health 3 (2024) 100108

[82] R.C. Caroline, Cost-effectiveness of robotics and artificial intelligence in health [86] C. Kooli, H. Al Muftah, Artificial intelligence in healthcare: a comprehensive
care, Univ. West. Ont. Med. J. 87 (2) (2019) 49–51, https://doi.org/10.5206/ review of its ethical concerns, Technol. Sustain. 1 (2) (2022) 121–131.
uwomj.v87i2.1179. [87] O. Ali, W. Abdelbaki, A. Shrestha, E. Elbasi, M.A. Alryalat, Y.K. Dwivedi,
[83] C. Robertson, A. Close, C. Fraser, et al., Relative effectiveness of robot-assisted and A systematic literature review of artificial intelligence in the healthcare sector:
standard laparoscopic prostatectomy as alternatives toopen radical prostatectomy benefits, challenges, methodologies, and functionalities, J. Innov. Knowl. 8 (1)
for treatment of localised prostate cancer: a systematic review and mixed treatment (2023) 100333.
comparison meta-analysis, BJU Int. 112 (6) (2013) 798–812, https://doi.org/ [88] S. Ali, F. Akhlaq, A.S. Imran, Z. Kastrati, S.M. Daudpota, M. Moosa, The
10.1111/bju.12247. enlightening role of explainable artificial intelligence in medical & healthcare
[84] K. Ahmed, A. Ibrahim, T.T. Wang, et al., Assessing the cost-effectiveness of robotics domains: a systematic literature review, Comput. Biol. Med. (2023) 107555.
in urological surgery: a systematic review, BJU Int. 110 (10) (2012) 1544–1556. [89] D. Lee, S.N. Yoon, Application of artificial intelligence-based technologies in the
[85] D.B. Olawade, O.J. Wada, A.C. David-Olawade, E. Kunonga, O.J. Abaire, Ling, healthcare industry: opportunities and challenges, Int. J. Environ. Res. Public
Using artificial intelligence to improve public health: a narrative review, Front. Health 18 (1) (2021) 271.
Public Health 11 (2023) 1196397. [90] S. Chatterjee, S. Das, K. Ganguly, D. Mandal, Advancements in robotic surgery:
innovations, challenges and future prospects, J. Robot. Surg. 18 (1) (2024) 28.

View publication stats